alterlab-genai-talking-head
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions are entirely focused on creative production workflows and professional persona definition. No patterns related to prompt injection, obfuscation, persistence, or privilege escalation were detected.
- [INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection as it is designed to ingest and process external data, though no malicious intent is present.
- Ingestion points: The agent context ingests user-provided photos, audio files, scripts, and brand guidelines from the local filesystem, as well as platform updates from the web.
- Boundary markers: The instructions do not define specific delimiters or 'ignore' commands for instructions embedded within the processed content.
- Capability inventory: The agent has the ability to search the web and write structured production guides and identity cards as markdown files to the local directory.
- Sanitization: There are no instructions for sanitizing or validating external input before processing.
- Contextual Mitigation: The skill includes explicit ethical constraints, such as mandatory disclosure of AI-generated content and likeness protection, which align with safe operational practices.
Audit Metadata